Hang around Internet discussions long enough and you’ll see history repeat itself, over and over. One very predictable event that takes place when I argue Internet policy is something I call the "gotcha weasel attack," so I’ve prepared this stock message in response. It’s on my personal web site at http://bennett.com/weasel_attack.html. The "gotcha weasel" has lost an argument with me, and sought to salvage his position by declaring me a paid shill of some commercial interest or other. He’s cited my think tank employment and offered all sorts of totally uninformed speculations about the nature of think tanks, their sponsors, and the flexibility of my opinions. He’s very proud of the mad black hat hacker skillz he’s displayed by doing a Google search on my name - I always use my real name on the Internet and I’m hit 3 and 4 on Google - and finding my resume. He’s declared that my opinions are not to be trusted, despite the factual or analytical evidence I’ve provided to back them up. If he’s extremely clever, he’s discovered that my employer doesn’t provide a list of sponsors, and demanded that I disclose them. The enormous irony of all this is that the gotcha weasel is probably using a fake name or anonymous handle, so the kind of personal research he’s done on me can’t be turned around. He’s hoping to derail the discussion by taking it down a rathole of personal attack and defense. I’m all for openness and transparency, hence I make no effort to hide my identity or my current or past employment. A diligent search of my body of work (or even a cursory glance at my resume) will reveal a consistent stand regarding the philosophy of network architecture dating back to the early 80s, when I was the vice chair of the IEEE 802.3 1BASE5 task force that formulated the first standard for Ethernet over UTP. I brought this philosophy to bear on my work on OSI and IETF standards in the mid-80s, when I formed the Open Token Foundation in the late 80s, and when I contributed to the Wi-Fi architecture and protocols from their foundation in 1990 up to 802.11n. At the end of this page are some links to things I wrote about networking before my current think tank stint. The approach I favor takes into account certain key facts about networking: 1.    Because networks are shared systems, every user affects other users, but users don’t have the means to effectively coordinate with others because they lack information about what others are doing. 2.    The most efficient networks use the information that can only be harvested at network aggregation points such as Ethernet switches, Wi-Fi access points, and Internet Exchanges to coordinate network use. 3.    Applications have different requirements, and well-functioning networks strive to produce the greatest good for the greatest number. 4.    Quality of Service is important in networks that support a diverse range of applications. 5.    The best form of congestion management is economic: It’s necessary to levy a higher fee for using a congested resource than an uncongested one. 6.    Networks don’t manage themselves, and it’s a poor network that doesn’t incorporate diagnostics into its essential design. 7.    The Internet is not a network; it’s a virtual network overlay that depends on the good graces of real networks to function. 8.    The larger the network, the more difficult it is to change, either for the better or for the worse. 9.    The proper way to build networks is neither to centralize nor to de-centralize; it’s to aggregate and disaggregate in hierarchical fashion. This method permits policy-driven use of shared resources according to whatever policy is appropriate to the resource in question. 10.  Nothing in the world is scarier than watching law professors try to design networks, resource sharing systems, and network regulations. Following ‘well established practices’ does not lead to technical progress. I’ve written dozens of posts on net neutrality since the debate started in the American media in 2006, and yet another dozen on Internet regulation before the public debate started. Most of my recent writing has been reacting to press reports, political events, and other people’s blog posts, and it’s fairly hard to follow, I expect, to those who haven’t been reading all along. So I decided to to collect the relevant considerations into a single post. Main points:

• Everything we know about regulating networks we learned from telephony.
• The Internet is radically different from the telephone network, hence traditional regulatory models don’t fit.
• The Internet is in its infancy and more experimentation is needed.
• Any regulation that’s not guided by empirical evidence of specific harm (not simply speculative, “what if?” scenarios) is likely to be wrong.
• The technical challenges to keeping the Internet running are so great that we don’t have the luxury of adding reams of unnecessary regulations to it.
• The appropriate regulatory stance is to watch for marketplace harm and be prepared to react to it.

These are some of my better historical posts on the subject: Net Neutrality Is Intelligent Design for the Left Quick note to Sen. Boxer Symmetry, Control, and Progress The Trouble With End-to-End How Much Bandwidth is Enough? Toward an Accountable Internet. And the complete archive is here. I hope this clears things up so we can get back to the flaming.